Adsorption of Enantiomers on Metal Surfaces

G102

Journal of The Electrochemical Society, 160 (8) G102-G110 (2013) 0013-4651/2013/160(8)/G102/9/$31.00 © The Electrochemical Society

Adsorption of Enantiomers on Metal Surfaces: Application to D- and L-Alanine on Cu, Ni and Zn Electrodes S. Harinipriya,a,z V. Sudha,b M. V. Sangaranarayanan,c and E. J. Padma Malard a Centre of Excellence in System Science, Indian Institute of Technology, Jodhpur, Rajasthan 342 b Department of Chemistry, Adhiyamaan College of Engineering, Hosur 635109, India c Department of Chemistry, Indian Institute of Technology, Madras, Tamilnadu 600 036, India d Department

011, India

of Physical Chemistry, University of Madras, Chennai 600 025, India

Different techniques have been developed for enantiomeric separation in order to meet the need for optically pure materials in the pharmaceutical, fine chemical and electronic industries. The present study explores the extent of selective adsorption of chiral compounds on metal electrodes, from knowledge of adsorption energy difference between D- and L- enantiomers. An entirely new simulation strategy is employed via Monte-Carlo method to evaluate the adsorption energy difference between D- and Lenantiomers. This methodology also yields the amount of each species adsorbed for a chosen electrode potential. The adsorption of tetrameric D- and L- alanines at Cu, Ni and Zn electrodes as well as in solution are studied using their stabilization energies obtained at the B3LYP/6-31G optimized structures. Subsequently these stabilization energies are employed as input parameters to estimate the adsorption energy difference between D- and L-alanine tetramers. The adsorption energy difference obtained from the simulation is found to be identical with the umbrella inversion energy for the lone pair of electrons on the amino group. It is demonstrated that, in a racemic mixture, only the D–alanine tetramer gets adsorbed predominantly on Cu, Ni and Zn while the adsorption of the L–species is more facile than the D-form when the corresponding pure enantiomer is employed. The origin of the preferential adsorption of D-enantiomer from a racemic mixture is interpreted using the computation of the molar volumes of the optimized geometries. Thus the evaluation of the adsorption energy of chiral compounds on metal electrodes can lead to valuable predictions for separation of optically active pure enantiomers. © 2013 The Electrochemical Society. [DOI: 10.1149/2.062308jes] All rights reserved. Manuscript submitted March 5, 2013; revised manuscript received April 17, 2013. Published May 18, 2013.

Perpetual growth in annual demand for enantiomerically pure chiral compounds occurs in the chemical, biochemical and pharmaceutical industries,1 the main focus being the study of (i) enantiospecific heterogeneous catalysis, bio-functionality, bio-toxicity effects in biochemical phenomena and (ii) pharmaceuticals development and manufacturing, associated with different enantiomers in vivo. Initial interest in enantiospecific heterogeneous catalysis is traced to the study of Orito reaction,2 carried out nearly two decades ago, wherein cinchonidine was employed as a chiral modifier for platinum surfaces in ± ketoester hydrogenation reactions. Subsequently the above reaction is applied to different reactants and modifiers.3 An alternate approach in the development of technology for enantiospecific production of useful compounds is based on the chemistry of the metal catalyst itself, rather than the reactants and modifiers chosen. The examination and initial nomenclature suggestions for chiral surface sites on metal surfaces such as the chiral kinks associated with the Pt(643) surface may offer the key to enantiomeric selection in adsorbates. It was the pioneering work of Sholl4 employing various crystal faces of Pt such as Pt(111) and Pt(643), on the behavior of dimethylcyclopropane and limonene, through umbrella-sampled Monte Carlo (MC) simulation techniques that pointed out the feasibility of enantiomeric selection in adsorption and desorption studies. The results are in agreement with the analysis of Gellman et al.5 which indicated that there should be appreciable experimentally detectable differences in the enantiospecificity of the surface in the Temperature Programmed Desorption (TPD), at least for some compounds. Attard et al.6 provided complementary results pertaining to the theory of enantiospecificity on naturally chiral metal surfaces. The analysis of glucose adsorbed on a variety of Pt single crystal faces as working electrodes for voltammetric study led to large, clearly discernable results in the differing permutations of glucose enantiomers and surface chirality for a given crystal face. These results yielded a different nomenclature system for the chiral surfaces, a refinement on the projections made by Gellman et al. This work along with the corresponding nomenclature protocol is still dominant in the area of chiral metal surface chemistry. Hovarth and Gellman7 observed that propylene oxide and methylcyclohexane underwent adsorption – desorption phenomena on chiral copper surfaces Cu(111) and Cu(643). Once again, for certain perz

E-mail: [email protected]

mutations of binding of species such as R–3–methylcyclohexane on Cu(643) surface, there are observable enantiospecific effects associated with the kink site. As mentioned earlier, another scenario of interest involves chiral molecules adsorbed on achiral surfaces such as the (100), (110), or (111) face of an fcc crystal, for example glycine on Cu(001) surfaces.8 It is also possible to induce chirality effects in pro-chiral molecules such as glycine adsorbed on achiral lattice faces. It was observed that the rate of deposition of adsorbate is a critical driving force in dictating the magnitude and nature of surface rearrangement. Apart from glycine, interfacial behavior of other aminoacids such as alanine, aspargine, aspartic acid etc have also been studied. Zhao et al.8 have investigated the adsorption behavior of amino acids on Cu(100) (or Cu(001 according to8 ). Amino acids with large alkyl groups characterized by carboxyl (L-aspargine) or carbonyl-amine (L-aspartic acid) do not organize into assembled patterns and domains on the surface. In contrast, other amino acids such as tryptophan exhibit organized assembly during adsorption. A typical analysis by Zhao et al. demonstrated the self-assembly of L-tryptophan on Cu (100) using Scanning Tunneling Microscopy (STM). The carboxyl group of the L-tryptophanate ions as well as the indole group bond to the Cu (100) surface with different monolayer coverages. The driving force for this behavior is attributed to hydrogen bonding between the ions, as well as π- stacking in parallel indole groups. A systematic study of adsorption of amino acids on different single crystals is especially crucial since (i) most organic molecules hitherto studied are either planar or rigid in contrast to amino acids which are flexible and chiral, possessing several active functional groups thus enabling us to explore the influence of different intermolecular and molecule-substrate interactions on self-assembly; (ii) proteins consisting of amino acids constitute nearly 50% of the dry mass of cells and hence understanding the adsorption mechanism of amino acids on inorganic surfaces is necessary for the development of biocompatible materials and (iii) proteins can impose order on mineral phases so as to produce the remarkable properties of bones, teeth and shells. For developing new materials, it is therefore imperative to commence the investigation of adsorption of amino acids on inorganic single crystal surfaces from an entirely new perspective. The advance of computing power and technology in recent years has allowed the application of Density Functional Theory (DFT) to larger systems, in more realistic timescales. This advance has prompted the collaboration of

Downloaded on 2013-05-20 to IP 117.211.84.218 address. Redistribution subject to ECS license or copyright; see ecsdl.org/site/terms_use

Journal of The Electrochemical Society, 160 (8) G102-G110 (2013) experimental and theoretical work at simultaneous or nearly simultaneous time frames. It is now well known that DFT in conjunction with experimental studies can play a valuable role in comprehending the adsorption behavior of enantiomers on various substrates. In the present work, the stabilization energies of the tetrameric structures of D- and L- alanine at Cu, Ni and Zn electrodes as well as at the solution are studied with the help of molecular dynamics simulation using DFT calculations at B3LYP/6-31G level. A new simulation strategy is proposed for estimating the adsorption energy difference between D- and L-configurations using the stabilization energies as the input parameters. The present simulation technique also yields the amount of each species adsorbed for a chosen electrode potential. In order to emphasis on the possibility of cost effective ways to separate enantiomers from racemic mixture via electrochemical techniques, polycrystalline metal surfaces such as Cu, Ni, Zn are considered in the present simulation rather than chiral kinks like Pt (643). Salient Features of the Methodology Molecular Dynamics (MD) and Monte Carlo (MC) simulations.— The MD simulation study for binding energies makes use of Hartree – Fock – Density Functional Theory (DFT) results obtained with Becke’s three-parameter hybrid-exchange functional and the gradientcorrected non-local correlation functional of Lee, Yang and Parr (B3LYP),9 using the Gaussian 03 software.10 All electron calculations using the split valence basis set were performed at B3LYP/6-31G level. The floppy molecules under study exist in shallow potentials and extensive computer time is needed for structural optimization due to very slow convergence. We imposed stringent convergence criteria using SCF = TIGHT option, in order to achieve Self-ConsistentField (SCF) convergence.11 Because of the large amount of computer time required to arrive at the optimized geometries, the basis set for structural optimization was restricted at the split valence level 6-31G, although 6-31G* basis set which includes polarization correction is known to be more accurate.11 However, single point B3LYP/631G*//B3LYP/6-31G calculations were done at the B3LYP/6-31G optimized geometries. The alanine molecules at the center of the cubic box were assumed to jump to a distance of 0.5 Å, on account of the applied potential (φapp ), along with its hydrogen bonded pairs and move progressively to the surface of the cube (representing the metal surface) and adsorb on the surface of the metal. The simulation is employed for obtaining (i) the number of molecules that arrive at the surface of the cube (metal) and (ii) the energy required for adsorption of each configuration employing the energy ratio as the criterion while generating random numbers for the analysis. From these values we arrive at the adsorption energy difference between the D- and L- species as shown below. The essential input parameters required for the present analysis are (i) the stabilization energies of the tetrameric D- and L- alanine in the bulk and at the metal electrode obtained from the DFT calculations and (ii) the conformation energy of the corresponding species on the metal surface. Using these two parameters, an explicit expression for the total energy required for the molecules to get adsorbed on the metal surface and their electrochemical potential at the metal surface are formulated. Before the simulation, the molecules are allowed to equilibrate for 0.03 pico sec, by rigidly allowing the molecules to move a particular distance (d) in the time ratio texp /teq , see eqn. 5. Simulation Details Molecular dynamics simulation.— Since the lowest energy conformations are the main focus of the study, we generated the Dand L-alanine tetramer conformations with the maximum number of intermolecular H-bonds. Starting from the optimized monomer conformation (both L- and D-), the second, third and fourth alanine units were oriented such that maximum number of H-bonds exist between the alanines. These conformations were then subjected to complete

G103

structural optimization at B3LYP/6-31G level. The suitability of the DFT method to yield reliable predictions on H-bonding interactions is established in earlier studies.12–18 Since the starting structures are chosen with maximum H-bonding interactions, the resulting optimized conformations are near the global minima. With the intermolecular H-bonding fixed, the conformational freedom in adopting different possible orientations with reference to the constituent alanine units is curtailed. At the optimized (D-ala)4 and (L-ala)4 geometries, the two metal atoms (M = Cu, Ni and Zn, as the case may be) were added and once again subjected to complete structural optimization. At the optimized geometries, we performed natural bond orbital (NBO) analysis19–21 and the covalent interactions were examined using covalent bond orders.22 Vibrational frequencies were calculated at the B3LYP/6-31G optimized geometries to ascertain their true minima status. No imaginary vibrational frequencies were present and thus the optimized structures were confirmed to be true minima in the potential energy surface. Zero-point vibrational energies (ZPE) were scaled by a factor of 0.9614 which was found suitable for B3LYP/631G* calculations.23 Single point energy calculations were carried out at the B3LYP/6-31G*//B3LYP/6-31G level by making use of the B3LYP/6-31G optimized geometries. The total energies and the zeropoint vibrational energies are presented in Table S1 in the Supporting Information. Stabilization energy of a given system E is obtained by subtracting the total energies of the components from the total energy of the system as shown below: E[(D − ala)4 ] = E[(D − ala)4 ] − 4 × E(D − ala)

[1]

E[(L − ala)4 ] = E[(L − ala)4 ] − 4 × E(L − ala)

[2]

E[(D−ala)4 M2 ] = E[(D−ala)4 M2 ]−4×E(D−ala)−E(M2 )

[3]

E[(L−ala)4 M2 ] = E[(L−ala)4 M2 ]−4×E(L−ala)−E(M2 )

[4]

Monte Carlo simulation.— The simulation was carried out in the NTP ensemble and the system consisting of the alanine molecules was placed in a cubic box of length lÅ. The simulation is performed for the chosen number density (molecules/cm3 ) of alanine. Boundary conditions were employed and the cube was confined (rigidly fixed) along z – axis. The molecules were initially placed at the center of the cube where it was equilibrated for 0.03 pico seconds whereas the experimental time for the molecules to feel the applied potential is 1 sec (vide infra) and allowed to move in x and y direction depending upon the initial displacement of the molecule. The distance (dtotal ) traveled by the molecules will be half the length of the cube (dtotal = l/2). Choice of input parameters for the Monte Carlo simulation.— (i)

The mean displacement of each alanine molecule in every 0.03 pico secs is d. Initially, before applying external potential φapp , d is assumed to be zero (the molecules are at the center of the cube). For every value of the applied potential (φapp ), d is calculated from an empirical expression d = texp ∗ x/teq

[5]

where ‘x’ denotes the expected displacement of the molecules (here assumed as half of the hydrogen bonding distance (0.5 A0 ) since in each displacement of the molecules the hydrogen bond is broken and formed simultaneously), texp being the experimental time of one sec24 and teq is the equilibration time during simulation. The ratio texp /teq will provide the equilibration of the structures when the potential φapp is applied. Here the expected displacement is assumed to be half the hydrogen bonding distance, since the only driving force for the tetramer structures is the hydrogen bonding between the four individual molecules.


G104

Journal of The Electrochemical Society, 160 (8) G102-G110 (2013) The hydrogen bonding distance is divided by two since during the simulation the alanine tetramers are allowed to move in either of the two directions x and y. While d is the expected displacement for each alanine tetramer from its original position, the actual displacement ‘d’ will be larger than this on account of the hydrogen–bonding between the four alanine molecules. Hence we write d = d H B + d

(ii)

[6]

where d H B is the hydrogen bonding distance (assumed to be equal to 1 A0 in the simulation; although the hydrogen bonding distance may extend to 2 A0 , the latter value when employed leads to ∼ 5% difference in computed energies). The energy (E mis ) involved in the displacement of the tetramer from its original position to a new position inside the cubic box depends on (i) the total distance covered by the species for reaching the metal surface from its new position and (ii) the expected displacement on account of the applied potential. Hence, E mis can be written as d ∗ φapp [7] E mis = dtotalr em where dtotalr em is the remaining distance the molecules have to travel from their new position to reach the metal surface (surface of the cube) subsequent to applying the potential φapp . dtotalr em is expressed as dtotalr em = dtotal − d

(iii)

the total energy involved in the orientation and the displacement of the tetramer from its initial position to the new position can now be written as ∗ E D−Ala dtotalr em + E mis Ei D = [9] d Ei L =

(iv)

[8]

∗ dtotalr em E L−Ala d

+ E mis

[10]

where E D−Ala and E L−Ala are the stabilization energies for the tetramers of D- and L- alanine in bulk, respectively, for the chosen metal surface. The electrochemical potentials involved in the adsorption process are represented as μ D = E i D − φapp

[11]

μ L = E i L − φapp

[12]

the adsorption energies of D- and L- alanine on the metal surface. However, for obtaining the number of D- or L- alanine molecules starting from a racemic mixture, the following energy criterion is employed: −(E D−ala + E total D ) [16] ir ( j1) >= exp E total D ir ( j2) >= exp

−(E L−ala + E total L ) E total L

[17]

If the generated random number satisfies eqn. 16, then D-alanine adsorbs on the metal surface, E total D being the adsorption energy of the D-alanine molecules obtained from eqns. 13 and 14. On the other hand, if the generated random number satisfies eqn. 17, then L-Alanine gets adsorbed, E total L being the adsorption energy of Lalanine on the metal. Scheme 1 indicates the steps involved in the simulation methodology. Results and Discussion Molecular dynamics simulation.— Hydrogen bonding interactions in the tetrameric clusters of D-alanine and L – alanine.—The fully optimized geometries of the tetrameric alanines shown in Figure 1 reveal the formation of three intermolecular H-bonds between the electropositive amino hydrogen of one of the D-alanines with the acceptor oxygen atom of neighboring D-alanine within the units D-ala1 , D-ala2 and D-ala3 . Further, these structures indicate the existence of a strong interaction between the nitrogen of D-ala3 and the carboxylic hydrogen of D-ala4 . Eventually the hydrogen atom H4 is pulled toward the nitrogen of D-ala3 , N3 , leading to covalent bond formation with bond length of 1.092 Å, which is about 0.08 Å longer than the B3LYP/6-31G predicted N-H bond length in alanine. Similarly, N4 of D-ala4 forms a covalent bond with the carboxylic hydrogen of D-ala3 . Consequently the bond lengths between O3 and H3 as well as O4 and H4 are elongated to 1.568 and 1.534 Å respectively and are characteristic of H-bond formation. Thus there are 5 intermolecular H-bonds in the cluster of (D-ala)4 . The lengths of these H-bonds lie in the range 1.53–2.12 Å and are shown in Figure 1

where μ D and μ L refer to the electrochemical potential for Dand L-species respectively. Simulation methodology.— In order to obtain the number of alanine molecules that reach the surface of the cube (equivalently the metal surface) random numbers are generated employing energy criterion as follows: ir ( j) > exp(−(E i L /E i D ))

[13]

Eqn. 13 yields the number of D- or L-alanine molecules reaching the metal surface. From the number of alanine molecules that reach the metal surface, the total energy required for their displacement from the center of the cube to the surface follows as E icon f D = (Nsur /N D−Ala )∗ E i D

[14]

E icon f L = (Nsur /N L−Ala )∗ E i L

[15]

and where Nsur denotes the number of molecules reaching the surface while ND-Ala and NL-Ala refer to the number of D- and L-alanine molecules after each displacement. E icon f D and E icon f L indicate respectively

Figure 1. B3LYP/6-31G optimized geometries of (D-ala)4 and (L-ala)4 . Labeling of atoms and the alanine units followed in the present study are also depicted.


Journal of The Electrochemical Society, 160 (8) G102-G110 (2013)

G105

Equilibrated alanine molecules at the center of the cube Movement of alanine molecules in x and y Computation of the distance ( d total ) traveled by the molecules ( d total = l / 2 )

The expected displacement of alanine molecules Δd = x * (

t exp t eq

)

The actual displacement‘ d ’,given by d = d HB + Δd ; where d HB is the hydrogen bonding distance (= 1A0).

Computation of the energy ( E mis ) involved in the displacement of the tetramer from its original position to a new position inside the cubic box as well as estimation of the remaining distance to be covered.;

⎛ Δd * φ app E mis = ⎜⎜ ⎝ d totalrem

⎞ ⎟ d totalrem = d total − d ⎟ ⎠

The total energy involved in the orientation and the displacement of the tetramer from its initial * d totalrem ⎞ * d totalrem ⎞ ⎛E ⎛E position to the new position E iD = E mis + ⎜ D − Ala ⎟ and E iL = E mis + ⎜ L − Ala ⎟; d d ⎝ ⎠ ⎝ ⎠ E D − Ala and E L − Ala are the stabilization energies for the tetramers of D and L alanine (computed for a Scheme 1. Estimation of the adsorption energy difference between D and L alanine molecules.

chosen metal surface) obtained from molecular dynamics simulations.

The electrochemical potentials involved in the adsorption defined as μ D = E iD − φ app and

μ L = E iL − φ app for D and L species respectively. Random number criterion ir ( j ) >= exp(−( E iL / E iD )) to obtain the total number of D and L alanine molecules reaching the metal surface from the bulk

The total energy leading to adsorption E iconfD = ( N sur / N D − Ala ) * E iD and E iconfL = ( N sur / N L − Ala ) * E iL . Nsur is the number of molecules that reached surface and ND-Ala

and NL-Ala are the number of molecules in each displacement of the alanine molecules. The adsorption energy difference between D and L alanine molecule ΔE ads = E iconfD − E iconfL

In the case of racemic mixture, ir ( j1) >= exp⎛⎜ − ( E D − ala + E totalD ) ⎞⎟ ⎟ ⎜ ⎝

(condition 1) and

⎛ − ( E L − ala + E totalL ) ⎞ ir ( j 2) >= exp⎜⎜ ⎟⎟ E totalL ⎝ ⎠

E totalD

⎠

(condition 2)

If the random number satisfies condition 1, then D-alanine gets adsorbed on the metal surface E totalD being the adsorption energy of the D-alanine on the metal, whereas if the random number satisfies condition 2 then L-Alanine adsorbs on the metal, E totalL being is the adsorption energy of the pure L-Alanine.

as broken lines. The bond angles D-H. . . A for these intermolecular H-bonding interactions are found to be 155–169◦ (Table I). The directionality and the lengths of these H-bonds indicate moderately strong H-bonds. The formation of covalent bonds N3 -H4 and N4 -H3 causes considerable structural reorganization in the residues D-ala3 and D-ala4 . This reorganization leads to the formation of intramolec-

ular H-bonds N3 H . . . O3 in D-ala3 and N4 H . . . O4 in D-ala4 with lengths of 2.02 and 1.93 Å respectively. These two H-bonds have bent-structures with bond angles of 109.5◦ and 113.9◦ , indicating that they are weak bonds. The stabilization energies listed in Table III shows that the hydrogen bonding interactions in the tetrameric Dalanine cluster increases the stability of the system by 34.0 kcal/mol


G106


Table I. Directionality of the hydrogen bond formation in the B3LYP/6-31G optimized geometries of D-alanine tetramer and its bimetallic complexes: The hydrogen bond angles D-H. . . A are given in degrees.

Table III. Stabilization energies (kcal/mol) of D-alanine and Lalanine in the bulk as well as on the bimetallic complexes of Ni, Cu and Zn. Parameter

H-bond angle

(D-ala)4

(D-ala)4 Ni2

(D-ala)4 Cu2

(D-ala)4 Zn2

N1 -H. . . O2 N2 -H. . . O1 N2 -H . . . O3 O3 . . . H3 -N4 O4 . . . H4 -N3 N3 -H . . . O3 N4 -H . . . O4

168.6 154.5 161.2 156.0 158.7 109.5 113.9

169.4 151.9 160.8 157.2 165.6 105.9 115.8

170.1 151.7 161.5 161.7 162.1 110.0 112.8

170.3 154.4 164.2 162.8 115.0 116.0 109.3

relative to the four non-interacting D- alanine monomers, according to B3LYP/6-31G level calculations with ZPE correction. The stabilization energy is predicted to be −39.6 kcal/mol when ZPE correction is not taken into account. The results are consistent with the presence of 7 H-bonds in the cluster indicating that the average H-bond energy is 4.9–5.7 kcal/mol. However, B3LYP/6-31G*// B3LYP/6-31G predict a smaller stabilization energy of −21.7 kcal/mol. The calculations suggest that in the tetramer of L-alanine, the geometry of the individual alanine molecules are not altered significantly. From Figure 1, it is inferred that three intermolecular H-bonds contribute predominantly to the stability of the L-tetramer. The lengths and angles of these H-bonds are about 2.15 Å and 150◦ (Table II). It is seen from Table III, that the energy of the tetramer is lowered by 8 kcal/mol as compared to the energy of 4 non-interacting alanines, indicating that the stability per H-bond is about 2.7 kcal/mol, according to B3LYP/6-31G calculations with ZPE correction. The cluster is predicted to be marginally more stable (8.4 kal/mol) at B3LYP/6-31G*// B3LYP/6-31G level. Analysis of the optimized geometries of bimetallic complexes of (D-ala)4 cluster (Figure 2) indicates that the metal centers are oriented closer to the carboxylic group of the D-ala4 unit. Thus the alanine units 1, 2 and 3 do not undergo any significant structural change from that of the tetrameric cluster (D-ala)4 . All the H-bonding interactions present in (D-ala)4 are retained in the complexes (D-ala)4 M2 with maximum deviation of about 0.04 Å in the H-bond lengths and 5◦ in the H-bond angles (Table I), in general. Somewhat larger deviations in the H-bonding parameters are observed when D-ala4 is involved, due to the proximity of the metal centers. The metal atoms exhibit significant covalent interactions with the oxygen and carbon centers of the carboxylic group of D-ala4 since they are located closer to them, particularly in the Ni and Cu complexes. The significant structural parameters indicating the interactions around the metal centers are listed in Table IV. The covalent bond orders are shown inside parenthesis. The calculations predict that a strong covalent interaction exists between Ni2 and O2 leading to bond formation with Ni2 -O2 bond length of 1.738 Å, which is typical of Ni-O single bond. Similarly a single bond is formed between Ni1 and C3 having the bond length of 1.816 Å and a bond order of 0.745. It is clear from Table IV that the interactions between the metal atoms with D-ala4 decrease in the order Ni > Cu > Zn. In the Ni com-

M1 -M2 M1 . . . O4 M2 . . . O4 M2 . . . C3 M1 . . . C3 M1 . . . C1 M1 . . . HC O4 . . . H4 N3 -H4 N4 -H3 O3 . . . H3 C1 C2 C3 C1 M1 M2 M1 M2 O4 C1 C2 C3 M2

(D-ala)4

1.534(0.163) 1.092(0.579) 1.083(0.594) 1.568(0.149) 112.2

(D-ala)4 Ni2

(D-ala)4 Cu2

(D-ala)4 Zn2

2.143(1.055) 2.647 (0.156) 1.738 (0.635) 2.387 (0.113) 1.816 (0.745) 2.456 (0.073) 1.851 (0.093) 1.502 (0.189) 1.113 (0.549) 1.099 (0.559) 1.488 (0.177) 107.1 131.1 82.5 87.0

2.151 (0.941) 2.008 (0.123) 1.997 (0.169) 2.011(0.398) 2.053 (0.365) 3.484 2.728 1.647 (0.123) 1.074 (0.614) 1.102 (0.557) 1.480 (0.181) 114.7 80.5 92.6 61.9

2.592 (0.185) 2.221 (.056) 2.176 (0.062) 2.904 2.958 5.015 4.687 4.003 1.095 (.554) 1.100 (.564) 1.515 (0.177) 112.2 83.4 79.6 144.9

plex, in addition to the interactions of Ni atoms with the carboxylic group, Ni1 interacts with the methyl group of D-ala4 which is reflected by the shorter distances, viz. Ni1 . . . C1 = 2.456 Å and Ni1 . . . Hc = 1.851 Å. These interactions are absent in the Cu and Zn complexes since there exists a large separation between the metals and the methyl group.

Table II. Directionality of hydrogen bond formation in the B3LYP/6-31G optimized geometries of tetramer of L-alanine and its bimetallic complexes: The hydrogen bond angles D-H. . . A are given in degrees. H-bond angles

N1 -H. . . O2 N2 -H. . . O3 N3 -H. . . O4 N3 -H. . . O4 N3 -H . . . O2

(L-ala)4 150.2 149.1 150.5

(L-ala)4 Ni2

(L-ala)4 Cu2

(L-ala)4 Zn2

149.2 135.9 146.9

149.5 134.4 139.2

147.3 153.5 169.7 143.8

Figure 2. B3LYP/6-31G optimized geometries of (D-ala)4 M2 where M refers to Ni, Cu or Zn.



G107

Table IV. Selected structural parameters in the B3LYP/6-31G optimized geometries of tetramer of D-alanine and its bimetallic complexes: bond lengths in Å, bond angles and dihedral angles in degrees. Covalent bond orders are given inside parenthesis. B3LYP/6-31G corrected for ZPE

B3LYP/6-31G

B3LYP/6-31G*// B3LYP/6-31G

(D-ala)4 (L-ala)4 (D-ala)4 Ni2 (L-ala)4 Ni2− (D-ala)4 Cu2 (L-ala)4 Cu2 (D-ala)4 Zn2 (L-ala)4 Zn2

−34.0 −8.0 −120.3 −157.8 −80.5 −136.9 −56.8 −72.5

−39.6 −10.0 −125.3 −157.8 −85.8 −138.2 −62.5 −73.6

−21.7 −8.4 −105.8 −154.7 −68.9 −132. 3 −46.3 −65.5


B3LYP/6-31G corrected for ZPE −34.0 −8.0 −120.3 −157.8 −80.5 −136.9 −56.8 −72.5

B3LYP/6-31G −39.6 −10.0 −125.3 −157.8 −85.8 −138.2 −62.5 −73.6

B3LYP/6-31G*// B3LYP/6-31G −21.7 −8.4 −105.8 −154.7 −68.9 −132. 3 −46.3 −65.5

The metal – metal bond lengths are predicted to be 2.143, 2.151 and 2.592 Å respectively for the Ni-Ni, Cu-Cu and Zn-Zn bonds in the complexes. Although the calculated bond orders show that the M1 -M2 (M = Ni, Cu) bonds are single bonds, the Zn-Zn bonding is considerably weak with a long bond length of 2.592 Å and a small bond order (0.185). This may be due to the closed shell d10 configuration of Zn. Table V shows that the interactions of the Zn atoms with D-ala4 is considerably lower than those of the Ni and Cu in the bimetallic complexes, as reflected by the larger distances of separation. On electrode surfaces, (D-ala)4 gets stabilized due to the interaction with the metal. According to the B3LYP/6-31G calculations, the stabilization energies in (D-ala)4 M2 are −120.3, −80.5 and −56.8 kcal/mol respectively for M = Ni, Cu and Zn, when ZPE correction is added (Table III). This lowering in the magnitude of stabilization energy is as anticipated. Thus the increase in the stabilization energy subsequent to adsorption is ∼ 86, 46 and 23 kcal/mol, respectively, for the complexes of Ni, Cu and Zn at the three different levels of calculation (Table III).

Figure 3. B3LYP/6-31G optimized geometries of (L-ala)4 M2 where M refers to Ni, Cu or Zn.

Interactions in (L-ala)4 M2 (M = Ni, Cu, Zn).—From the B3LYP/631G optimized structures of these complexes depicted in Figure 3, it is clear that the orientations of the metal atoms favor interactions with the carboxylic group and the nitrogen center N4 of the L-ala4 . Unlike in the corresponding D-analogs in which the metal atoms interact only with D-ala4 , it is seen that in the L-complexes the metal atom M2 is also proximal for interaction with the carboxylic oxygen O3 of the L-ala3 . Thus structural parameters in the residues L-ala3 and L-ala4 undergo considerable reorganization. The major changes that occur in

Table V. Selected structural parameters in the B3LYP/6-31G optimized geometries of tetramer of L-alanine and its bimetallic complexes: bond lengths in Å, bond angles and dihedral angles in degrees. Covalent bond orders are given inside parenthesis. Parameter M1 -M2 M1 . . . N4 M1 . . . O4 M1 . . . O4 M2 . . . O4 M2 . . . O3 M2 . . . C3 M1 . . . C3 M1 . . . H4 M2 . . . H4 O4 . . . H4 C1 C2 C3 C1 M1 M2 M1 M2 O4 C1 C2 C3 M2 C1 M1 M2 O4

(L-ala)4

0.982 (0.150) 111.4

(L-ala)4 Ni2

(L-ala)4 Cu2

(L-ala)4 Zn2

2.108 (0.874) 2.244 (0.123) 1.904 (0.408) 2.780 1.889 (0.349) 1.988 (0.118) 2.303 (0.148) 1.952 (0.390) 1.679 (0.333) 1.711 (0.314) 3.363 113.8 78.2 88.0 38.8 −50.7

2.225 (0.413) 1.926 (0.233) 3.111 2.968 1.943 (0.152) 2.061 (0.094) 2.658 (0.063 2.588 (0.058) 1.592 (0.384) 1.605 (0.298) 4.049 115.0 80.8 90.6 45.9 −47.0

2.383(0.507) 2.129(0.074) 4.237 2.175 2.119(0.062) 2.140(0.057) 3.360 3.017 3.730 1.558(0.519) 4.858 112.4 109.6 57.4 171.5 17.4


G108


Table VI. The number of D- and L- species that adsorb on the metal surfaces and their variation with the stabilization energies and nature of the metal surface. Applied potentials are 0.001 V, 0.01 V and 0.1 V. Number of alanine tetramers molecules that adsorbs on the metal surface

M Ni Cu Zi Ni Cu Zi Ni Cu Zi

Number of alanine tetramers and box size 100 and 10 Å

100 and 15 Å

150 and 10 Å

B3LYP/6-31G with ZPE correction

B3LYP/6-31G

B3LYP/6-31G*//B3LYP/6-31G

D

L

D

L

D

L

1.3 2.3 1.3 1.3 2.3 1.3 1.8 3.7 1.8

14 24.9 14 14 24.9 14 20 40.4 20

1.2 2.0 0.8 1.2 2.0 1.2 1.8 3.5 1.2

12.7 21.8 8.6 12.7 21.8 12.7 19.3 38.3 13.4

1.7 3.3 1.7 1.7 3.3 1.7 2.8 5.3 2.7

18.5 36.1 18.6 18.5 36.1 18.5 29.8 58.3 29.7

the structural parameters surrounding the metal atoms are provided in Table V. Figure 3 shows that the positions of the metals are favorable for bonding with oxygen atoms of the carboxylic group of L-ala4 . As observed in the case of (D-ala)4 M2 complexes, interactions between the two metal centers and the tetrameric cluster of L-alanine follows the decreasing order for Ni > Cu > Zn. This is evident from the shorter distances of 1.904, 1.889 and 1.952 Å, respectively, for Ni1 . . . O4, Ni2 . . . O4 and Ni1 . . . C3 . In the case of the Cu and Zn complexes, these distances are longer and the covalent interactions are less pronounced (Table V). Another striking observation in the (L-ala)4 M2 complexes is that the interactions of the metal atoms are able to push the carboxylic H4 of L-ala4 such that it is detached (O4 -H4 >3.3 Å) and is located closer to the metal atoms from the opposite side (Figure 3). The calculations reveal that in the complexes [(L-ala)4 M2 , M = Ni, Cu], the detached H4 interacts with both M1 and M2 resulting in partial bond formation with bond orders in the range 0.30–0.38. However in the corresponding Zn complex, H4 adopts an orientation that favors interaction with Zn2 only. The covalent bond formation between the Zn2 and H4 is significant with a bond order of 0.52. The L-ala4 unit is twisted in the zinc complex and O4 comes nearer to both the zinc atoms with a separation of 2.175 and 2.119 Å. Consequently no H-bonds exist between O4 and the amino hydrogen of L-ala3 . Instead, O4 takes part in the H-bond formation. Also, in the case of Zn complex, an additional H-bond is formed between O2 of L-ala2 with hydrogen of the amino group in L-ala3 . In view of the interactions between the metal atoms with L-ala3 and L-ala4 , the H-bonds between L-ala2 and L-ala3 as well as that between L-ala3 and L-ala4 undergo changes. There is weakening of the H-bond between L-ala2 and L-ala3 as reflected by the increase in the H-bond length to 2.25–2.30 Å in the metal complexes as compared to the value of 2.15 Å in (L-ala)4 . The H-bond angle for the above bonds decreases by ∼15◦ in the Ni and Cu complexes but the angle gets widened to 153.5◦ in the Zn complex (Table II). As the interactions with the metal centers increase the negative charges on the oxygen atoms of L-ala4 , the H-bond between L-ala3 and L-ala4 is strengthened further and the H-bond length is decreased by ca. 0.15 Å. The H-bond between L-ala1 and L-ala2 is however, not affected in the metal complexes. Although the L-tetramer of alanine is energetically less stable than the D-tetramer (Table III), the present study shows that in the bimetallic complex, the two metal atoms lead to stronger stabilizing interactions with the electronegative centers in L-ala3 and L-ala4 units than in the D- analogue. This is evident from a comparison of distances and bond orders collected in Tables IV and V. Further, some of the H-bonding interactions have become stronger in the L-complex due to an increase in negative charge on oxygen center of L-ala4 on account of the influence of the metal. Thus, for example, the bonding with the two nickel centers has stabilized (L-ala)4 by 146–150 kcal/mol at different levels of computation. The corresponding stability in the

case of (D-ala)4 Ni2 is about 84–86 kcal/mol. Thus the overall stability of the complex (L-ala)4 M2 has increased as compared with the corresponding (D-ala)4 M2 counterpart. The relative energies listed in Table III reveals that the increased stabilities of the bimetallic complexes of the L-tetramer over (D-ala)4 M2 are in the ranges 33–49, 52–63 and 11–19 kcal/mol, respectively, for M = Ni, Cu and Zn at the different levels of calculation. Monte Carlo simulation studies.— The Monte Carlo simulation yields the adsorption energy difference between D- and L- alanine at a chosen applied potential and identification of the predominant configuration (viz D or L) getting adsorbed in the case of a racemic mixture. Three sets of data were obtained at (i) different applied potentials (0.001 V, 0.01 V and 0.1 V) (ii) varying sizes of the cubic box (10 Å and 15 Å) and (iii) different number densities (100 and 150). Although the number of molecules is ∼ 102 , they were chosen in order to demonstrate the adsorption energy differences even when low concentrations are employed. We have simulated the adsorption of the pure (D-ala)4 as well as (L-ala)4 enantiomers on the metal electrodes by separately taking 100 (or 150) molecules inside a cubic box of length 10 Å (or 15 Å). Table VI summarizes the results obtained. Tables S3 to S29 of the Supporting Information provide the adsorption energy difference for each step movement of the D- and L- species from Monte Carlo simulation. It is inferred from these data that neither the applied potential nor the number of alanine molecules alters E ads but the nature of the metal influences its value. From Table VI, it is seen that the adsorption of the D-species is within 4% under the different conditions while there is about a 10-fold increase in the adsorption of the L-enantiomer. The exact number of D- and L-species getting adsorbed is dependent upon the input stabilization energies at different levels (cf. B3LYP/6-31G with and without ZPE correction or B3LYP/6-31G*//B3LYP/6-31G). It is evident that the increase in the adsorption of the L- alanine tetramer is due to the increase in the stability of the metal complex (L-ala)4 M2 formed (Table III). Further, it is observed that the number of L-species adsorbed also depends on the magnitude of the relative stabilization between the L- and the D-analogs. The DFT calculations using the polarized basis set at the B3LYP/6-31G*//B3LYP/6-31G level predict increased relative stabilities as compared to B3LYP/631G calculations with and without ZPE correction. The present Monte Carlo simulation reveals that the number of (L-ala)4 species adsorbed on a given metal electrode is highest at B3LYP/6-31G*//B3LYP/631G level of calculation. Another interesting observation is that the tendency for D- species to get adsorbed is more pronounced when a racemic mixture is employed as shown in Table VI. Adsorption of alanine tetramer on nickel.—For a fixed number of 100 and 150 D-enantiomer molecules chosen initially, the simulation indicates that ca.1 to 3 molecules get adsorbed on the nickel surface. On the other hand, for 100 and 150 (L-ala)4 molecules present inside



G109

Table VII. The adsorption energy difference at an applied potential of 0.001 V in the case of Nickel; dtotal = 10 A0 and number of alanine molecules = 100.

No. of steps 1 2 3 4 5 6 Number of D or L species reaching the metal surface

E ads on Nickel obtained by simulation employing the computed stabilization energies from B3LYP/6-31G B3LYP/6-31G B3LYP/6-31G*//Corrected to ZPE B3LYP/6-31G 0.00033954

0.00028892

0.00046936

0.00067281 0.0010 0.0013 0.0017 0.0020 1.3 D–Alanine

0.00057309 0.00086164 0.0011 0.0014 0.0017 1.2 D–Alanine

0.00093234 0.0014 0.0019 0.0023 0.0028 1.7 D–Alanine

14 L–Alanine

12.7 L–Alanine

18.5 L–Alanine

the cube, 19 and 29 molecules are adsorbed on the Ni surface, respectively, at the B3LYP/6-31G*// B3LYP/6-31G level (Table VI). This observation reveals that adsorption of the L-alanine tetramer is more facile than the D-analog and this behavior is attributed to additional stabilization energy for the L-alanine tetramer-Ni2 complex, which is 48.9 kcal/mol more than that in the D-complex. The number of (L-ala)4 adsorbed on the nickel surface is predicted to be about 14% and 13% respectively at B3LYP/6-31G level with and without ZPE correction, as expected from the decreased relative stabilities of 38 and 33 kcal/mol. Adsorption of alanine tetramer on copper.—Table VI shows that the adsorption of alanine tetramer on copper electrode follows analogous trend as in the case of Ni electrode. However, the number of L-species adsorbed is significantly higher and ranges from 21 to 36 (when 100 alanine molecules are initially assumed) or 38 to 58 (for 150 initial molecules) whereas the corresponding number of D -alanine tetramer adsorbed is roughly 10% of the above value. The significant increase in the adsorption of L-species is due to the large stabilization energy difference between the L- and D- alanine tetramers on copper surface. As seen from Table III, the stabilization energy of L-alanine tetramerCu2 complex is nearly twice that of the corresponding D-counterpart. Adsorption of alanine tetramer on zinc.—In this case too, the number of L - alanine tetramer molecules getting adsorbed is ca. 10 times larger than the D- species as shown in Table VI. It is noticed that the number of D- and L- molecules adsorbed on zinc surface is nearly same as that on nickel surface, although the stabilization energy difference between the L- and D-alanine tetramer-Zn2 complex is small (Table III). Adsorption of racemic mixture of alanine tetramer on metal electrodes.—When a racemic mixture is used for the Monte Carlo simulation study, the present methodology indicates that the adsorption of D-alanine tetramer occurs rather than L-alanine on the metal surfaces (M = Ni, Cu, Zn) (cf. Table VII, as an illustrative example the behavior of racemic mixture on Ni (conditions being 0.001 V, 10A0 and 100 molecules). It is seen from the Supporting Information provided that the adsorption energy difference, E ads , follows the

Inference Only D – Alanine is adsorbed on the metal

The% of D : L on the surface is ca. 1:10

sequence: E ads(Cu) > E ads(N i) > E ads(Z n) E ads values on Copper, Nickel and Zinc are estimated as 0.002 eV, 0.004 eV and 0.00084 eV respectively. Since E ads = E icon f 1 − E icon f 2 , where E icon f 1 and E icon f 2 represent the adsorption energies of D- and L- species respectively, it is clear that the D- alanine (rather than L- alanine) from a racemic mixture gets adsorbed strongly on the metals in the order Copper > Nickel > Zinc. The estimated E ads values on Copper, Nickel and Zinc exactly match with the umbrella inversion energy of the lone pair of electrons on the amino group of the amino acids. Thus it is deduced that the umbrella inversion governs the orientation and adsorption of D- alanine tetramer molecules on metal surfaces such as Cu, Ni and Zn. This observation is rationalized on the basis of the strain experienced by the L - alanine tetramers for attaining a favorable conformation with respect to the metal surface vis a vis the competition between Dand L- species for adsorption. From the optimized structures shown in Figure 1, it may be inferred that the L-alanine tetramer is bulkier than the D-analog, which has more stabilizing H-bonding interactions and thus possesses a compact structure. This is quantified from the molar volumes computed at the B3LYP/6-31G optimized geometries using the Gaussian software.10 It is clear from Table VIII that (D-ala)4 is 18.6 cc/mol smaller in volume than that of (L-ala)4. Thus the approach of (D-ala)4 from the racemic mixture to the metal surface is less hindered as compared to that of the L-tetramer. A comparison of the molar volumes of the metal complexes in Table VIII also reveals that the D-enantiomer is smaller in size than the L-counterpart. It is striking to note that the difference in volumes between the Land D- complexes follows the order Cu (28.374) > Ni (16.388) > Zn (2.832). This feature again substantiates the predicted adsorption energy differences. The CPU time utilized for Monte Carlo simulations are reported in Table IX. The stabilization energies (Table III) used as input parameters in the present MD simulation analysis were generated by DFT calculations in vacuum. Although solvents are expected to play important role in the stabilization energies and structures of molecules in

Table IX. CPU time for Monte Carlo Simulations. Table VIII. Molar volume (in cc/mol) of the D- and L-enantiomers. System

Molar volume, cc/mol


253.702 272.152 281.468 297.856 280.947 309.321 310.294 313.126

System

CPU time (hrs)

100 molecules of alanine (Cu, Ni and Zn metals), length of the box 10 A0 150 molecules of alanine (Cu, Ni and Zn metals), length of the box 10 A0 100 molecules of alanine (Cu, Ni, Zn metals), length of the box 15 A0

3 hrs for each metal and each applied potential 4.5 hrs for each metal and each applied potential 5.5 hrs for each metal and each applied potential


G110


general, our calculations for the systems under study in water medium using the Onsager model,25–27 at the B3LYP/6-31G optimized geometries in vacuum, show only minor changes in the total energies and in the stabilization energies (Table S1 in the supporting information). The difference in stabilization energies in water as well as in vacuum for tetrameric L-alanine amounts to 1.1 kcal/mol. Further, the D- and L-complexes of Zn show variations of 1.2 and 2.4 kcal/mol respectively (Table S1). In the remaining systems studied, the difference in the stabilization energies in water and vacuum is much less than 1 kcal/mol. This clearly indicates that the interactions between the solvent water and the solute molecules under study are very small. Thus it is reasonable to assume that the structure in the aqueous solution is very close to that of the corresponding vacuum optimized structure. Thus, although the role of solvent is not included herein, the insights obtained from the present analysis are expected to be valid in solutions also. Summary The stabilization energies of tetrameric structures of D- and Lalanine molecules at the as well as at Cu, Ni and Zn electrodes were studied using molecular dynamics simulation at B3LYP/6-31G level. These stabilization energies were employed as the input parameters in estimating the adsorption energy difference between D- and Lalanine tetrameric molecules using a novel simulation methodology. This approach, which invokes the energy ratio as the criterion, yields the adsorption energy difference between D- and L- alanine tetramic molecules on Cu, Ni and Zn. This energy difference was found to be consistent with the Umbrella inversion energy for lone pair of electrons on the nitrogen of the amino group. The new simulation technique is demonstrated to provide (i) the amount of each configuration getting adsorbed for a chosen electrode potential and (ii) the identification of the configuration in the case of a racemic mixture. The Monte Carlo simulation in the present investigation is different from the classical MC techniques where the potential is defined by Lennard Jones or other electrostatic interactions (potential truncation and long range interactions). Instead in the current Monte Carlo simulation process the stabilization energies evaluated from molecular dynamic simulations are fed as the input energies to generate random numbers in Monte Carlo simulation. This coupling of Molecular Dynamic simulation with Monte Carlo simulation enables handling of higher number of molecules with less tedious computation efforts.

References 1. A. L. Lehninger, L. D. Nelson, and M. M. Cox, Principles of Biochemistry, 2nd edn. 1993, CBS publishers and references therein. 2. P. N. Rylander, Catalytic Hydrogenation over Platinum Metals, NY, Academic Press, 1967. 3. S. Lavoie, M. A. Laliberte, and P. H. McBreen, J. Am. Chem. Soc., 125, 15756 (2003). 4. S. D. Sholl, Langmuir, 14, 862 (1998). 5. K. R. Paserba and A. J. Gellman, J. Chem. Phys., 115, 6737 (2001). 6. G. Attard, J. Phys. Chem. B, 105, 3158 (2001). 7. J. D. Horvath and A. J. Gellman, J. Am. Chem. Soc., 123, 7953 (2001). 8. X. Zhao, Hao Yan, R. G. Zhao, and W. S. Yang, Langmuir, 19, 809 (2003). 9. A. D. Becke, J. Chem. Phys., 98, 5648 (1993); C. Lee, W. Yang, and R. G. Parr, Phys. Rev. B, 37, 785 (1988). 10. Gaussian 03, Revision B.03, M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, J. A. Montgomery Jr., T. Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar, J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega, G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratchian, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, P. Y. Ayala, K. Morokuma, G. A. Voth, P. Salvador, J. J. Dannenberg, V. G. Zakrzewski, S. Dapprich, A. D. Daniels, M. C. Strain, O. Farkas, D. K. Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul, S. Clifford, J. Cioslowski, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W. Chen, M. W. Wong, C. Gonzalez, and J. A. Pople, Gaussian, Inc., Wallingford CT, 2004. 11. J. B. Foresman and A. Frisch, Exploring Chemistry with Electronic Srtucture Methods; Gaussian, Inc.: Pittsburgh, PA, (1996). 12. C. L. Perrin and J. B. Nielson, Annu. Rev. Phys. Chem., 48, 511 (1997). 13. K. M¨uller-Dethlefs and P. Hobza, Chem. Rev., 100, 143 (2000). 14. K. M. Specht, J. Nam, D. M. Ho, N. Berova, R. K. Kondru, D. N. Beratan, P. Wipf, R. A. Pascal Jr., and D. I. Kahne, J Am. Chem. Soc., 123, 8961 (2001). 15. J. E. Del Bene, Hydrogen bonding: 1. Encyclopedia of Computational Chemistry; D. Schleyer, Editor in Chief; John Wiley: Chichester, UK, 1998; Vo1. 2. 16. J. E. Del Bene and M. J. T. Jordan, J. Mol. Struct. (THEOCHEM), 573, 11 (2001). 17. J. Ireta, J. Neugebauer, and M. Scheffler, J. Phys. Chem. A, 108, 5692 (2004). 18. J. He and P. L. Polavarapu, J. Chem. Theory Comput., 1, 506 (2005). 19. J. P. Foster and F. Weinhold, J. Am. Chem. Soc., 102, 7211 (1980). 20. A. E. Reed, R. B. Weinstock, and F. Weinhold, J. Chem. Phys., 83, 735 (1985). 21. A. E. Reed, L. A. Curtiss, and F. Weinhold, Chem. Rev., 88, 899 (1988). 22. K. B. Wiberg, Tetrahedron, 24, 1083 (1968). 23. A. P. Scott and L. Radom, J. Phys. Chem., 100, 16502 (1996). 24. The experimental time of one second is realistic in a potential sweep experiment wherein t = (E − E i )/ν where E and E i refer respectively to the applied potential range and the initial potential. The scan rate is denoted by ν. For values of (E − E i ) ranging from 0.001 V to 0.1 V with a scan rate of 1 mV per sec to 1000 mV per sec, the experimental time may range from 1 second to 1000 seconds. 25. L. Onsager, J. Am. Chem. Soc., 58, 1486 (1936). 26. M. W. Wong, M. J. Frisch, and K. B. Wiberg, J. Am. Chem. Soc., 113, 4776 (1991). 27. M. W. Wong, K. B. Wiberg, and M. J. Frisch, J. Chem. Phys., 95, 8991 (1991).
